3/31/2026
Multi-Agent Actor-Critics in Autonomous Cyber Defense
Multi-Agent Deep Reinforcement Learning with Actor-Critic algorithms applied to autonomous cyber defense mechanisms.
3/31/2026
Efficient Human-in-the-Loop Active Learning: A Novel Framework for Data Labeling in AI Systems
Active learning framework optimizes expert time for labeled data collection in AI systems, demonstrated on medical imaging.
3/31/2026
Gradient Compression Beyond Low-Rank: Wavelet Subspaces Compact Optimizer States
Wavelet subspace gradient compression reduces LLM training memory requirements for optimizer states beyond low-rank techniques.
3/31/2026
Hybrid Action Based Reinforcement Learning for Multi-Objective Compatible Autonomous Driving
Hybrid action reinforcement learning achieves multi-objective compatibility in autonomous driving decision-making.
3/31/2026
A Survey of Zero-Knowledge Proof Based Verifiable Machine Learning
Survey of zero-knowledge proof based verifiable machine learning for ensuring computational integrity and privacy in ML pipelines.
3/31/2026
Shifting Perspectives: Steering Vectors for Robust Bias Mitigation in LLMs
Applies steering vectors to modify LLM activations for bias mitigation across social bias axes including gender, age, race.
3/31/2026
Towards Quantifying Long-Range Interactions in Graph Machine Learning: a Large Graph Dataset and a Measurement
Large graph dataset and measurement approach quantify long-range interactions in graph neural networks versus message-passing models.
3/31/2026
Benchmarking NLP-supported Language Sample Analysis for Swiss Children's Speech
NLP methods analyze Swiss children's speech for developmental language disorder diagnosis without commercial LLM dependency.
3/31/2026
Measuring the (Un)Faithfulness of Concept-Based Explanations
Measures faithfulness of concept-based explanation methods for deep vision models by evaluating surrogate accuracy.
3/31/2026
BRIDGE: Benchmarking Large Language Models for Understanding Real-world Clinical Practice Text
Benchmark evaluating LLMs on real-world electronic health record clinical text understanding across diverse medical models.
3/31/2026
Beyond Recognition: Evaluating Visual Perspective Taking in Vision Language Models
Evaluates vision-language models on visual perspective taking using controlled scenes with humanoid figures and objects.
3/31/2026
Self-Bootstrapping Automated Program Repair: Using LLMs to Generate and Evaluate Synthetic Training Data for Bug Repair
LLMs generate synthetic training data to improve automated program repair systems across diverse bug types and programming languages.
3/31/2026
FlowPure: Continuous Normalizing Flows for Adversarial Purification
Continuous normalizing flows remove adversarial perturbations from ML models at inference time as defense strategy.
3/31/2026
Structured Agent Distillation for Large Language Model
Framework compressing large LLM-based agents into smaller student models while preserving reasoning and action fidelity using distillation.
3/31/2026
VLM-SAFE: Vision-Language Model-Guided Safety-Aware Reinforcement Learning with World Models for Autonomous Driving
Vision-language models guide safe reinforcement learning for autonomous driving using world models to reduce unsafe trial-and-error.
3/31/2026
Learning to Diagnose Privately: DP-Powered LLMs for Radiology Report Classification
LLMs with differential privacy applied to radiology report classification for multi-abnormality detection in healthcare workflows.
3/31/2026
Can Generalist Vision Language Models (VLMs) Rival Specialist Medical VLMs? Benchmarking and Strategic Insights
Benchmarking study comparing generalist and specialist vision-language models for medical image interpretation with insights on when each excels.
3/31/2026
Multi-Sample Prompting and Actor-Critic Prompt Optimization for Diverse Synthetic Data Generation
Multi-sample prompting with actor-critic optimization for improving diversity and quality of LLM-generated synthetic data for machine learning.
3/31/2026
MicroMix: Efficient Mixed-Precision Quantization with Microscaling Formats for Large Language Models
MicroMix enables efficient mixed-precision quantization for LLMs using microscaling formats and FP4 Tensor Cores for faster inference.
3/31/2026
AirQA: A Comprehensive QA Dataset for AI Research with Instance-Level Evaluation
AirQA dataset and benchmark for evaluating LLM-based agents on scientific paper question-answering with instance-level evaluation metrics.
3/31/2026
Multi-View Attention Multiple-Instance Learning Enhanced by LLM Reasoning for Cognitive Distortion Detection
Combines LLMs with multiple-instance learning for cognitive distortion detection in mental health applications, enhancing interpretability and reasoning.
3/31/2026
Dual-Space Smoothness for Robust and Balanced LLM Unlearning
Dual-space smoothness approach for machine unlearning in LLMs balancing unlearning effectiveness, utility preservation, and privacy protection simultaneously.
3/31/2026
More Thought, Less Accuracy? On the Dual Nature of Reasoning in Vision-Language Models
Empirical study examining the trade-off between reasoning chain length and accuracy in vision-language models trained with reinforcement learning.
3/31/2026
Generating Findings for Jaw Cysts in Dental Panoramic Radiographs Using a GPT-Based VLM: A Preliminary Study on Building a Two-Stage Self-Correction Loop with Structured Output (SLSO) Framework
SLSO framework uses GPT vision-language models with self-correction loops and structured output for automated dental radiograph interpretation and medical image analysis.
3/31/2026
Agentic Context Engineering: Evolving Contexts for Self-Improving Language Models
ACE framework evolves contexts for self-improving language models, addressing brevity bias and context collapse through agentic context engineering for domain-specific applications.
3/31/2026
Efficient Tree-Structured Deep Research with Adaptive Resource Allocation
ParallelResearch framework enables efficient deep research agents through parallel tree-structured reasoning with adaptive resource allocation instead of sequential processing.
3/31/2026
Mitigating Premature Exploitation in Particle-based Monte Carlo for Inference-Time Scaling
Mitigates premature exploitation in particle filtering for inference-time scaling of language models using process reward models for improved mathematical reasoning.
3/31/2026
Open ASR Leaderboard: Towards Reproducible and Transparent Multilingual and Long-Form Speech Recognition Evaluation
Open ASR Leaderboard benchmarks 86 open-source and proprietary speech recognition systems across 12 datasets with standardized evaluation metrics for multilingual and long-form audio.
3/31/2026
Dream to Recall: Imagination-Guided Experience Retrieval for Memory-Persistent Vision-and-Language Navigation
Dream to Recall presents a memory mechanism for vision-and-language navigation agents using imagination-guided experience retrieval for progressive improvement.
3/31/2026
CLMN: Concept based Language Models via Neural Symbolic Reasoning
CLMN introduces concept bottleneck models for NLP with neural-symbolic reasoning to improve interpretability in language models by tying predictions to human concepts and modeling dynamic interactions.
3/31/2026
SceneAdapt: Scene-aware Adaptation of Human Motion Diffusion
Diffusion-based human motion generation adapting to scene constraints using two-stage approach for diverse, scene-aware motion.
3/31/2026
Schema for In-Context Learning
Schema-based in-context learning framework enabling transformers to activate and transfer pre-existing knowledge abstractions.
3/31/2026
Learning When Not to Learn: Risk-Sensitive Abstention in Bandits with Unbounded Rewards
Bandit learning algorithm with abstention mechanism for high-stakes applications requiring risk-sensitive decision-making.
3/31/2026
ProofBridge: Auto-Formalization of Natural Language Proofs in Lean via Joint Embeddings
Auto-formalization system translating natural language mathematical proofs to Lean 4 using joint embedding approach.
3/31/2026
BIOGEN: Evidence-Grounded Multi-Agent Reasoning Framework for Transcriptomic Interpretation in Antimicrobial Resistance
Multi-agent framework using evidence-grounded reasoning for transcriptomic interpretation in antimicrobial resistance research.
3/31/2026
DIV-Nav: Open-Vocabulary Spatial Relationships for Multi-Object Navigation
Robot navigation system interpreting open-vocabulary spatial relationship queries for multi-object search.
3/31/2026
Evaluating Latent Knowledge of Public Tabular Datasets in Large Language Models
Framework for detecting data contamination in tabular datasets within LLMs through latent knowledge assessment.
3/31/2026
Dense and Diverse Goal Coverage in Multi Goal Reinforcement Learning
Multi-goal reinforcement learning approach for dispersed state coverage while maximizing expected return.
3/31/2026
ViPRA: Video Prediction for Robot Actions
Framework learning robot control policies from unlabeled video data using video prediction and action inference.
3/31/2026
Beyond Elicitation: Provision-based Prompt Optimization for Knowledge-Intensive Tasks
Prompt optimization approach providing factual knowledge and terminology for knowledge-intensive LLM tasks beyond elicitation.
3/31/2026
$\pi$-Attention: Periodic Sparse Transformers for Efficient Long-Context Modeling
Periodic sparse attention mechanism reducing transformer complexity for long-context modeling with adaptable receptive fields.
3/31/2026
ImAgent: A Unified Multimodal Agent Framework for Test-Time Scalable Image Generation
ImAgent framework using AI agents to improve text-to-image generation consistency through iterative refinement without additional modules.
3/31/2026
Scaling Spatial Intelligence with Multimodal Foundation Models
Scaling multimodal foundation models for improved spatial understanding capabilities in visual reasoning tasks.
3/31/2026
Object-Centric World Models for Causality-Aware Reinforcement Learning
Object-centric world models for sample-efficient reinforcement learning by decomposing environments into discrete entities.
3/31/2026
SciEGQA: A Dataset for Scientific Evidence-Grounded Question Answering and Reasoning
Dataset for scientific document QA with explicit evidence annotation for improved interpretability and evaluation reliability.
3/31/2026
Towards Hyper-Efficient RAG Systems in VecDBs: Distributed Parallel Multi-Resolution Vector Search
Multi-resolution vector search optimization for RAG systems in vector databases to improve retrieval speed and contextual relevance.
3/31/2026
SAM 3: Segment Anything with Concepts
SAM 3 extends segment anything model to detect, segment and track objects using concept prompts combining text and image exemplars.
3/31/2026
UniGame: Turning a Unified Multimodal Model Into Its Own Adversary
Research on unified multimodal models addressing trade-offs between understanding and generation through adversarial training to improve cross-modal coherence.
3/31/2026
From Observation to Action: Latent Action-based Primitive Segmentation for VLA Pre-training in Industrial Settings
Unsupervised framework for VLA model pre-training using motion tokenization and action segmentation on industrial video.
3/31/2026
Single-Round Scalable Analytic Federated Learning
SAFLe: Single-round federated learning framework for non-linear models with heterogeneous data.